Clear Of The Low-visibility Image Processing

The visibility of images which are obtained in bad weather is low. The value of this image is decreased greatly, and they make bad effect on the image processing in traffic, military, aerospace etc. For this reason, the study on how to clear the images has very important significance.First, the influence of atmospheric effect on the image is present.The image restoration methods such as inverse filtering, wiener filtering and index model are analyzed. But building a model of atmosphere degradation is difficult, and the calculation of recovery algorithm is large. Then we find that the atmospheric degradation image has the characteristics of low brightness and low contrasr. As a result we use the image enhancement method to clear the image with low visibility.Then, we discuss and compare the historgram equalization, homomorphic filter and the image enhancement method based on Retinex etc. Because the SNR is low when the visibility of the image is low, we use the mothod of pre-processing and post-pocessing to improve the traditional algorithm. We statistics the SNR of the image. and adopt the image denoing if the SNR is substandard, And we use the method of linear stretching to improve the image which is enhanced. Simulation results show that the pre-treatment and post-treatment can effectively improve the visual effect of image enhancement.Finally, in order to adapt to the demand of machine vision, we study the POBC(the product of brightness and contrast) and VCM to measure the quality of image enhancement, and propose the adaptive MSR algorithm based on POBC and VCM. We classify the image with the POBC value, choose corresponding MSR parameters to enhance the image, then measure the POBC and VCM of the image which is improved, and the evalution result is feedback to adjust MSR parameter until the quality of image is up to standard. Simulation results show that this algorithm can further improve the MSR enhancement processing of visual characteristics.